agent-opfor

Open-source adversary emulation for AI agents, LLM apps, and MCP servers. Test your AI like a real attacker would — from your CLI, your IDE, or a browser extension that anyone on your team can use. Website · Docs · GitHub · Browser Extension · Discord OPFOR is short for Opposition Force — a...

aimap

AIMap Internet-scale discovery and security testing platform for exposed AI agent infrastructure. AIMap finds, fingerprints, and security-tests publicly exposed AI endpoints — MCP servers, Ollama instances, vLLM/LiteLLM proxies, LangServe chains, Gradio apps, ComfyUI nodes, and more. Think Shodan...

mcp-xray

MCP X-Ray !NOTE Looking to secure AI agent skills? Check out Skill X-Ray, an open-source static security scanner that analyzes skill packages for malicious or risky behavior before installation, without executing them. Overview MCP X-Ray is a unified open-source security scanning and penetration...

agent-scan

Snyk Agent Scan Discover and scan agent components on your machine for prompt injections and vulnerabilities including agents, MCP servers, skills. Note: We don't publish an npm package for Agent Scan. Install it via uvx or as a standalone binary. Note: CLI output is experimental and subject to...

CVE-2026-105761

The open-source LLM app development platform Dify contains an Insecure Direct Object Reference (IDOR) vulnerability in versions prior to 1.16.0 . The root cause is located in api/controllers/console/app/mcp_server.py, where the PUT /console/api/apps//server endpoint uses AppMCPServerController.pu...

CVE-2026-105697

Langflow is a tool for building and deploying AI-powered agents and workflows. Before Langflow 1.10.3, the MCP stdio transport launched whatever command / args a user put in an MCP server configuration, with no allowlist and before 1.10.3 wrapped in bash -c "exec command ...". Any user able to...

CVE-2026-105697

Langflow versions prior to 1.10.3 are vulnerable to OS command injection due to a lack of allowlisting in the MCP stdio transport configuration. An attacker capable of accessing MCP server settings or using the MCP Tools component can execute arbitrary OS commands (e.g., rm -rf, reverse shells) o...

CVE-2026-105697 Langflow: OS command injection (RCE) via arbitrary command in MCP stdio server configuration

Langflow is a tool for building and deploying AI-powered agents and workflows. Before Langflow 1.10.3, the MCP stdio transport launched whatever command / args a user put in an MCP server configuration, with no allowlist and before 1.10.3 wrapped in bash -c "exec command ...". Any user able to...

vulnerable-mcp-servers-lab

취약한 MCP 서버 랩 이 저장소는 의도적으로 취약하게 만든 Model Context Protocol MCP 서버로컬 및 원격 구현을 포함합니다. 각 서버는 자체 폴더에 있으며 기능 , 실행 방법 , 취약점 시연/공격 방법 에 대한 자세한 내용이 포함된 전용 README.md 파일을 제공합니다. 통제된 랩 환경 외부에서 실행하지 마십시오. 이 저장소의 용도 일반적인 MCP 서버 및 도구 통합 실패 모드에 대한 보안 교육/연구. 취약한 MCP 서버가 데이터 노출, 명령어 삽입, 공급망 손상, 코드 실행으로 이어질 수 있는 방법에...

CVE

CVE 표 CVE 번호| 유형| 업체/제품| 설명| 기사 링크| 저장소 링크 ---|---|---|---|---|--- CVE-2026-9680| SQLi| Alibaba RDS MCP Servers| TBA| TBA| TBA CVE-2026-50335| LPE| Windows July 2026| TBA| TBA| TBA GHSA-73cv-556c-w3g6| SQLi| Apache Pinot MCP 3.1.0| 외부에 바인딩된 MCP 서버의 입력 검증 누락으로 SQL 인젝션이 발생할 수 있음|...

LLM-MCP-Security-Field-Guide

🛡️ AI Security Field Guide — LLM + MCP Security The most comprehensive, up-to-date, practitioner-first security reference for LLM applications and Model Context Protocol MCP deployments. Covers real CVEs, live attack patterns, OWASP frameworks, red teaming tools, and actionable checklists — update...

pkgxray

pkgxray — npm 패키지, MCP 서버, AI 에이전트를 위한 설치 전 보안 npm 패키지나 MCP 서버를 설치하거나 연결하기 전에 검사하여 결정적이고 증거 기반의 SAFE, REVIEW, BLOCK 판정을 받으세요. 로컬, 제로 의존성 정적 분석 — 일반 스캔은 패키지 코드를 절대 실행하지 않습니다. 웹사이트 · 문서 · 캘리브레이션 · 버그 신고 실제 실행: guard가 [email protected]을 통과시킨 뒤, 2024년 @solana/web3.js 침해 사고를 모델링한 샘플을 차단합니다. 주요 기능...

skills

SpecterOps Skills Reusable agent skills, plugins, and agent definitions for SpecterOps. Use With Codex Each plugin includes a Codex manifest at .codex-plugin/plugin.json. codex plugin marketplace add /Users//Projects/skills or codex plugin marketplace add SpecterOps/skills Then open Codex and...

mcp-server-attestation

StudioMeyer MCP Stack의 일부 — Mallorca에서 제작 🌴 · 사용한다면 ⭐를 눌러주세요 mcp-server-attestation Model Context Protocol 서버를 위한 레이어-2 공급망 강화. Ed25519 서명 툴 매니페스트, 런타임 spawn 증명, 기본 거부default-deny 인자 샌니타이저. 다음에 대한 직접적인 대응: OX Security 마켓플레이스 오염, 2026년 4월 — MCP 레지스트리 11곳 중 9곳이 악성 서버를 수용했습니다. Anthropic이 공식적으로 밝힌 입장...

CVE-2026-103758

Obot 0.21.1 through 0.24.1 contains an authorization bypass vulnerability that allows authenticated users to reach MCP servers because the checkUI deny list omits the /mcp-connect-composite/ route. Basic-role users with a composite MCP ID can proxy requests through mcpGateway.Proxy to invoke tool...

CVE-2026-103758

The Obot platform, versions 0.21.1 through 0.24.1 , is vulnerable to an authorization bypass because the checkUI deny list fails to include the /mcp-connect-composite/ route. This flaw allows authenticated users with basic-role privileges and a composite MCP ID to proxy requests via mcpGateway.Pr...

EUVD-2026-90687

Obot 0.21.1 through 0.24.1 contains an authorization bypass vulnerability that allows authenticated users to reach MCP servers because the checkUI deny list omits the /mcp-connect-composite/ route. Basic-role users with a composite MCP ID can proxy requests through mcpGateway.Proxy to invoke tool...

CVE-2026-103758 Obot 0.21.1 through 0.24.1 Authorization Bypass via /mcp-connect-composite/ Route

Obot 0.21.1 through 0.24.1 contains an authorization bypass vulnerability that allows authenticated users to reach MCP servers because the checkUI deny list omits the /mcp-connect-composite/ route. Basic-role users with a composite MCP ID can proxy requests through mcpGateway.Proxy to invoke tool...

CVE-2026-103758: Incorrect Authorization

Obot 0.21.1 through 0.24.1 contains an authorization bypass vulnerability that allows authenticated users to reach MCP servers because the checkUI deny list omits the /mcp-connect-composite/ route. Basic-role users with a composite MCP ID can proxy requests through mcpGateway.Proxy to invoke tool...

CVE-2026-101084 obot before v0.21.1 Authorization Bypass via /mcp-connect

obot versions before v0.21.1 fail to enforce Access Control Rules on the /mcp-connect endpoint, allowing any authenticated user to connect to restricted MCP servers if they possess the server ID. Attackers can bypass authorization checks to access and manipulate sensitive backend systems through...